PREHISTORIC MODIFICATION OF THE FOREST COVER
IN THE DELAWARE COASTAL PLAIN:
AN INTERPRETIVE MODEL FOR LARGE D-SHAPED PIT FORMATION
CHRIS EGGHART
PARSONS
ABSTRACT
A well-known aspect of Delaware prehistoric archaeology is the occurrence of very large apparent
pit features identified on prehistoric site locations. These features, which often exhibit a “D”-shaped or
crescentic plan, have variously been interpreted as the remains of pit houses, simple tree falls, or as
resulting from the interactions of multiple formation processes. This paper presents an alternative
interpretation for the formation of these features. It is argued that these archaeological entities share a
common, cultural origin. It is also postulated that the features represent disturbances caused by the active
modification of the local forest cover by prehistoric peoples. Such modification would have been
undertaken in order to promote the growth of edible wild plant food species, enhance mast production by
eliminating competing trees, facilitate nascent horticulture, provide a ready supply of limb firewood, or a
combination of these impetuses.
INTRODUCTION
Large D-shaped features, commonly encountered on prehistoric sites in Delaware were first documented in significant numbers on a series of archaeological mitigations undertaken by the University of
Delaware Center for Archaeological Research in association with State Route 1 construction. Major sites
on which these features were documented include Leipsic (Custer et al. 1994a), Pollack (Custer et al.
1994b), Snapp (Custer and Hsiao Silber 1994) and Carey Farm and Island Farm (Custer et al. 1995) sites.
The very large features, upwards to three meters in size, are described as being D-shaped in plan, with
other examples having a crescentic or kidney-like configuration (Figure 1). Profiles often show one steep
side with the opposite flaring outwards at a shallower angle. Considerable controversy surrounds these
entities. In fact, most researchers who have worked in Delaware have been unable to agree on a functional
interpretation and even whether or not the features are cultural in origin. For the purposes of discussion,
these entities are collectively referred to as D-shaped pits.
D-shaped pits have been identified on additional major excavations undertaken in central and
northern Delaware by Parsons, including Hickory Bluff (Petraglia et al. 2002) Glasgow School (Bowen et
al. 2003) Frederick Lodge (Egghart et al. 2003), as well as on the Puncheon Run (LeeDecker et al. 2001)
complex investigated by Louis Berger and Associates.
THE DEGRADED PIT HOUSE MODEL OF D-SHAPED PIT FORMATION
Custer interpreted the D-shaped pits as the signature remains of semi-subterranean structures or “pit
houses”. According to this interpretation, the features represent storage facilities located inside shallow
83
84
Journal of Middle Atlantic Archaeology, Volume 21, 2005
Figure 1. D-shaped pit (from Petraglia et al. 2002).
semi-subterranean structures, most other evidence of which has been lost to plowing. The conceptual
template for this was based on a single finding made on the Snapp site in New Castle County (Custer and
Hsiao Silber 1994). Here a complex of features was encountered at the foot of a minor slope. The addition
of colluvial soil had protected this area from the extensive plow truncation evident across the rest of the
site. Designated Feature 153, the complex consisted of a narrow D-shaped pit set into one end of a much
larger, shallow basin. A third small pit was defined in the approximate center of the larger feature. The
cluster of pits was roughly enclosed by a pattern of nine, irregularly spaced post molds. The Feature 153
complex was interpreted as a dwelling composed of a shallow “basement”, or living area, with a “subbasement” storage facility located at one end with the post mold pattern representing a superstructure
constructed over the below-grade house floor (Custer and Hsiao Silber 1994:43-52). This interpretative
reconstruction was applied to a large number of the Snapp site D-shaped pits and extended to other
Delaware sites on which similar features were present (Custer et al. 1994b:36-38, Custer et al. 1995:2523). On the Leipsic site alone (Custer et al. 1994a), a total of 197 individual features were presented as
sub-basement storage facilities located within dwellings, all other traces of which have been lost to
plowing and soil deflation (Figure 2). This interpretation for D-shaped pit formation is herein referred to
as the Degraded Pit House Model.
While this pit house interpretation for D-shaped features retains its proponents, it has been
questioned by other researchers working in the state. Mueller and Cavallo (1995) argued that the features
primarily represent tree throws, while leaving open the possibility that these tree throws could have been
utilized or modified by aboriginal peoples. LeeDecker (personal communication 1998) viewed the Dshaped pits as representing house remains with skepticism noting the general lack of posts or other
recognizable feature types normally associated with structure locations. Mueller and Cavallo (1995) also
noted that both direct and indirect evidence for such intensive prehistoric settlement as would be signaled
by hundreds of D-shaped pits representing structure locations, is completely lacking in the greater region.
Equally problematic is the proposed storage function. The flaring form is poorly suited as a storage
facility, particularly inside a structure where every square foot of floor space represents a significant
investment in labor and material. Large cylindrical pits (Figure 3) interpreted as storage facilities have
been identified in considerable numbers on Early and Middle Woodland sites in Delaware (Thomas 1981,
D-SHAPED PIT FORMATION
85
Figure 2. Degraded pit house (from Custer 1994:57).
LeeDecker et al. 2001, Petraglia et al. 2002). These features, dubbed silo pits (Petraglia et al. 1998) based
on their form and presumed function, bear little resemblance to the trench-like configuration of the Dshaped pits.
Other researchers have suggested that the D-shaped pits represent simple tree throws.
Contemporary tree falls and their resultant ground disturbances are readily observable in wide variety of
Coastal Plain locations. These disturbances have a shallow, saucer-like form. Commonly observed tree
throws typically are D-shaped in plan. The straight side marks the pivot axis (or fold) at ground level,
while the rounded end is created by the lateral root fan splaying upwards opposite the side to which the
tree fell. What is clearly lacking in the typical tree throw, however, is the deep, trench-like gouge that is a
defining characteristic of the D-shaped pit.
ARCHAEOLOGICAL REVIEW OF D-SHAPED PITS
A review of the D-shaped pit findings in on numerous major sites in the Delaware Coastal Plain
suggests that these entities share common morphological attributes. These include a D-shaped plan view
suggestive of a tree fall location and a deep gouge-like profile atypical of a tree throw. A review of
archaeological attributes of exhibited by the D-shaped pits on Delaware sites indicated that these features
consistently exhibited evidence of cultural input (Custer et al. 1994a; Custer et al.1995; Custer and HsiaoSilber 1995). The compilation of radiocarbon dates (Table 1) obtained from numerous D-shaped pits
documented by the University of Delaware proved to be in agreement with the major occupation periods
for the respective sites. The Hickory Bluff site D-shaped pits, in various permutations, yielded occupation
period dates, contained elevated artifacts counts, and/or had elevated levels of elemental phosphorous
(Table 2). At the Frederick Lodge site (Egghart et al. 2003) a single large D-shaped pit was fully
investigated. Chemical analysis of the fill soil indicated significantly elevated levels of elemental
86
Journal of Middle Atlantic Archaeology, Volume 21, 2005
Figure 3. Delaware silo pit (from Petraglia et al. 1998).
TABLE 1: RADIOCARBON DATES FROM LARGE PIT FEATURES ON DELAWARE STATE
ROUTE 1 SITES
Snapp Site (7NC-G-101) (Source: Custer and Hsiao-Silber 1994:103)
Feature
Date B.P.
Date Calibrated
Notes
105
142/193
153
206
1410"70
1150"80
2420"70
1640"70
AD 576 – 666
AD 775 – 984
BC 350 – 72
AD 262 – 531
Carey Farm (7K-D-3) (Source: Custer, et al. 1995:146)
Feature
Date B.P.
Calibrated Date
623
1640"70
AD 370 – 530
371
1240"60
AD 695 – 880
427
1680"60
AD 535 – 635
440
1720"60
AD 245 – 410
465
1300"60
AD 665 – 785
608
1660"50
AD 370 – 435
686
1260"70
AD 680 – 875
Lab Number
D-shaped
D-shaped
Snapp House
D-shaped
Beta-56802
Beta-56361
Beta-56803
Beta-56360
Notes
D-shaped
D-shaped
D-shaped
D-shaped
D-shaped?
Round pit?
D-shaped
Lab Number
Beta-76845
Beta-76837
Beta-76838
Beta-76839
Beta-76840
Beta-76841
Beta-76842
D-SHAPED PIT FORMATION
Leipsic Site (7K-C-194A) (Source: Custer, et al.1994a: 69)
Feature
Date B.P.
Date Calibrated
353 Area A
3650"100
2192 - 1890 BC
314 Level 4
2070"70
199 BC – AD 1
255/256 A
1080"130
AD 778 – 114
255/256 D
770"170
AD 1040 –1390
77 Level 3
1820"110
AD 60 – 340
266
1400"80
AD 576 – 674
Notes
D-shaped
D-shaped
D-shaped
D-shaped
D-shaped
D-shaped
87
Lab Number
Beta-42878
Beta-42879
Beta-42881
Beta-42882
Beta-42883
Beta-42884
TABLE 2: HICKORY BLUFF (7K-C-411) D-SHAPED ATTRIBUTE SUMMARY (Source:
Petraglia et al. 2002)
Feature # Tree Rot
Morphology
2
No
77
No
90
No
118
No
169
No
Tree Throw
Morphology
Partial
Partial
Partial
Partial
Partial
Cultural
Morphology
No
No
No
No
No
Elevated
Artifacts #
No
No
Yes
Yes
No
Elevated
P Content
Yes
Yes
Yes
N/A
Yes
Absolute
Dating
2790"40 BP
N/A
4070"40 BP
N/A
N/A
phosphorous near the bottom of the pit, suggesting cultural input. At a minimum, these findings suggest
that the features were extant at the time of the occupations. While the data suggest a cultural association
or connection, the sheer numbers in which these features have been documented, and the repetition on
their basic form, both within and across sites, point toward a common formation process. While some
researchers maintain that the D-shaped pits resemble tree throws and have postulated that tree falls
constitute the primary mechanism for their formation, a significant body of data clearly points toward a
prehistoric cultural association. The following interpretive model reconciles these seemingly divergent
standpoints. The Culturally-Induced Tree Fall Model presents the D-shaped pit phenomenon as being
both the result of a tree fall and as having a cultural origin.
THE CULTURALLY INDUCED TREE FALL MODEL
The Culturally Induced Tree Fall model presents the D-shaped pits as constituting the physical scar
of a tree fall that had been induced to occur by prehistoric site occupants. The purpose of such activity
would have primarily been to open forest areas to promote the growth of food producing plant species,
and perhaps to enhance mast production by eliminating competing, non-economically valuable trees. A
key consideration is the fact that soils in the Delaware Coastal Plain are mostly sandy and unconsolidated,
allowing for easy excavation using primitive tools. Under these conditions, a tree could be induced to fall
by grubbing around the base of one side, leaving the roots in place. A fire constructed in the resultant
trench would burn through the exposed roots, either downing the tree outright, or killing it and leaving it
unstable and to be felled by the next storm. Figure 4 schematically illustrates the conjectural procedure, as
well as the anticipated archaeological signature. In this manner, a significant number of trees could be
brought down with a minimum of labor. The resulting dead fall could either be left in place or burned to
fully clear the area. This model can also explain a basic variability in D-shaped pits described by Custer
(1994). Should the tree fall in the direction of the grubbing trench, the extant root fan on the opposite
side would pull up, forming the typical saucer-like tree throw disturbance. This shallow depression would
be present on one side of the grubbing trench, as the roots on the opposite side would have been firesevered. Together the tree fall and the grubbing trench would leave a ground disturbance (Figure 5)
strikingly similar to the Custer’s (1994) Type 2 Delaware Woodland I feature type that forms the basis for
88
Journal of Middle Atlantic Archaeology, Volume 21, 2005
Figure 4. Culturally induced tree fall (from Egghart et al. 2003).
‘basement/sub-basement” pit house interpretation. Conversely, if the tree were to fall in the direction
away from the grubbing trench, then the lateral root fan on the “down side” side would simply fold over
and not generate a significant disturbance for the archaeological record. In this latter scenario, the primary
archaeological trace would be the grubbing trench excavated to expose the roots. Such a trench would
appear very similar to the more narrow, crescentic Type 1 pit as defined in Custer’s (1994) Woodland I
pit typology. Prehistoric grubbing and burning around the base of trees could also account for the
formation of the third type of very large pit categorized by Custer. The main characteristic of Type 6
feature is described as two opposing, crescentic trenches containing charcoal flecking (Custer 1994).
Grubbing to both sides, or fully around a standing tree is anticipated to result in just such a ground
disturbance as defined by Custer’s Type 6 feature.
D-SHAPED PIT FORMATION
89
Figure 5. Type 1 and 2 Woodland I pit features (from Custer 1994:58).
It is recognized that girdling a tree would be effective in creating forest clearings, and that both
Late Woodland horticulturists as well as early European settlers employed this technique. However, it is
the physical disturbance of the ground, as well as the interruption of the forest canopy from felling trees
that creates conditions most favorable to early succession plant communities. Selectively burning an area
also fully removes the understory, releases nutrients, and results in conditions favorable to natural
diversity (Sutton and Sutton 1984:24).
Peoples of the Delaware Early and Middle Woodland are generally regarded to have had a less than
fully sedentary lifeway, sustained by wild foods. Seasonal availability of these resources and other factors
dictated movements between occupation loci, probably within fairly regular and proscribed round. It is
suggested that within this settlement round, people opportunistically, but on an ongoing basis, modified
the local environment via the opening of minor forest tracts. This would have been done in order to
expand the productive “forest fringe” and enhance the growth of food-bearing plant species that thrive as
components of early succession communities in disturbed areas. The selective downing of trees may also
have constituted a form of “forest management” applied to increase mast production or to otherwise favor
individual economically important trees. Trees may also have been downed in order to provide an ample
supply of limb fuel for return visits during the course of regular seasonal settlement rounds. In this
context, it is useful to remember that prehistoric Native Americans were not capable of effectively
processing trunk firewood, in the modern sense, within the confines of a lithic tool technology. Ongoing
modification of the forest cover by prehistoric peoples would have had the ancillary effect of increasing
the game carrying capacity of the given area. Patchy woods in varying stages of succession, containing
open zones, breaks in the canopy and decaying trunks, are a more productive game habitat than a homogenous climax forest cover.
90
Journal of Middle Atlantic Archaeology, Volume 21, 2005
The interpretation of aboriginal clearing of forest areas to increase food bearing wild plant growth
dovetails with generally accepted local subsistence and settlement models for the period. These models
are predicated on continuing population growth and the advent of varying degrees of semi-sedentism,
factors that in turn beget intensification of food procurement/production. Some researches have used the
occurrences of the D-shaped pits along a site’s periphery or even beyond the prehistoric occupation loci,
to argue against a cultural formation of the features. In the Induced Tree Fall Model, this is precisely
where they would be expected to occur — not necessarily on the occupation site, but along its edges.
Paleo-Environmenal Data
Data generated from local paleo-environmental research sponsored by the Delaware Department of
Transportation in support of State Route 1 archaeological studies (Kellogg and Custer 1994) suggest that
an active modification of the local forest cover may have occurred during the Middle Woodland period.
Specifically, Brush’s (1994) analysis of St. Jones River sediment cores just downstream from Hickory
Bluff, and Cary Farm and Island Field sites show a dramatic, 400 percent spike in charcoal for samples
dating between 2000 and 1500 years. This spike in charcoal, and the stratigraphically congruent pollen
profile, are interpreted by Brush (1994:90-91) as evidence for frequent fires but with these events
occurring within an overall moisture regime not significantly different than today’s.
Regionally, there is a growing body of data suggesting an ongoing human alteration of the natural
environment occurred in Eastern North America beginning in the Late Archaic period. Delcourt et al.
(1986) provide data that suggest the floodplains of the Little Tennessee River were largely cleared by at
least 4000 B.P. Synthesizing broad regional data, Stevens (1991) argued that human intervention,
particularly the burning and clearing of floodplains, is reflected in geomorphologic record to include
increased stream sedimentation and localized aeolian soil deposition. The purpose of this intervention
would have to promote the growth of edible wild plant species that thrive as part of succession
communities. Bruce D. Smith (1989, 1992) has convincingly argued that the promotion of certain
pioneering, indigenous seed bearing species was undertaken in Eastern North America to the extent that
the region should be recognized as a locus for independent plant domestication. Smith postulates that the
intensive exploitation of indigenous seed-bearing plants developed into nascent horticultural practices
long before the introduction of tropical cultigens. The propagation of these pioneering species depends on
the physical disturbance of the ground as well as forest canopy. Smith (1989,1992) suggests that seasonal
flooding along the rivers in Eastern North America originally provided this ground disturbance and that
an exploitive focus of these areas developed into nascent horticultural practices a millennium or more
prior to the acceptance of tropical cultigens and development of fully horticultural lifeways. A prehistoric
modification of the local forest cover in the Delaware Coastal Plain is viewed within this context.
SUMMARY
In summation, it is hereby postulated that the distinctive D-shaped and crescentic pits, welldocumented on numerous sites in Delaware, by and large, represent the physical traces of prehistoric
peoples having purposefully induced trees to fall within and around their occupation loci. This would
have been accomplished by grubbing along the base of a mature tree, then burning through the exposed
roots. The primary purpose of this activity would have been to create forest openings for the growth and
propagation of food bearing wild plant species that thrive as components of early succession
communities. Trees may also have been downed along the periphery intensively occupied site areas and
left in place order to provide a ready supply of limb firewood for extended, and/or seasonally repeating
site visits. Minor tracts of forest may also have been selectively managed in order to increase mast
production or to favor the growth of individual economically useful trees. These actions would have been
undertaken within the context of a Woodland lifeway marked by diminishing residential mobility and an
intensification of local food resource exploitation.
D-SHAPED PIT FORMATION
91
REFERENCES CITED
Brush, Grace S.
1994
Paleobotanical Analyses of Three Tidal Stream Valleys Along the Proposed State Route 1
Corridor, Kent County Delaware. In Paleoenvironmental Studies of the State Route 1
Corridor: Contexts for Prehistoric Settlement, New Castle and Kent Counties, edited by
D.C. Kellogg and J.F. Custer, pp. 82-95. Delaware Department of Transportation
Archaeology Series No. 114, Dover.
Bowen, C., P. O’Neill and E.A. Crowell
2003
Archaeological Data Recovery Excavations at the Glasgow Elementary School Site, Site
7NC-D-212 William B. Keene Elementary School New Castle County Delaware. Parsons
Engineering Science, Fairfax, Virginia. Submitted to the Christina School District, New
Castle County, Delaware.
Custer, Jay F.
1994
Stability, Storage and Culture Change In Prehistoric Delaware: The Woodland I Period
(3000 B.C. – A.D. 1000). On file at the Delaware State Historic Preservation Office,
Division of Historical and Cultural Affairs, Dover.
Custer, Jay F., Lynn Riley and Glen Mellin
1994a
Final Archaeological Excavations at the Leipsic Site (7K-C-194A) State Route 1 Corridor,
Kent County, Delaware. University of Delaware Department of Anthropology Center for
Archaeological Research, Newark.
Custer, Jay F., A. Hoseth, B.H. Silber, D.J. Grettler, and Glen Mellin
1994b
Final Archaeological Investigations at the Pollack Site (7K-C-203), State Route 1
Corridor, Kent County, Delaware. Delaware Department of Transportation Archaeology
Series No. 126. Dover.
Custer, Jay.F., and B.H. Silber
1994
Final Archaeological Investigations at the Snapp Site (7NC-G-101), State Route 1
Corridor, Chesapeake and Delaware Canal Section, New Castle County, Delaware.
Delaware Department of Transportation Archaeology Series No. 122. Dover.
Custer, Jay F., Scott. C. Watson and B.H. Silber
1995
Final Archaeological Investigations at the Carey Farm (7K-D-3) and Island Farm (7K-C13) Sites, State Route 1 Corridor, Kent County, Delaware. Delaware Department of
Transportation Archaeology Series No. 146. Dover.
Delcourt, Paul A., Hazel R. Delcourt, P. A. Cridlebaugh, and Jefferson Chapman
1986
Holocene Ethnobotanical and Paleobotanical Record of Human Impact on Vegetation in the
Little Tennessee River Valley, Tennessee. Quaternary Research 25:330-349.
Egghart, Christopher P. and Dennis Knepper and Chris Bowen
2003
Prehistoric Settlement of Bay/Basin Landscapes in the Archaic and Woodland Periods:
Phase II and Phase III Archaeological Investigations at the Frederick Lodge Complex,
7NC-J-97, 7NC-J-98, and 7NC-J-99, New Castle County, Delaware. Delaware Department
of Transportation Archaeology Series, No. XXX (draft). Prepared for the Delaware
Department of Transportation, Dover.
Kellogg, Douglas C. and Jay F. Custer (editors)
1994
Paleoenvironmental Studies of the State Route 1 Corridor: Contexts for Prehistoric
Settlement, New Castle and Kent Counties. Delaware Department of Transportation
Archaeology Series No 114. Dover.
LeeDecker, C., J. Bedell and R. Jacoby.
2001
Archaeology of Puncheon Run Site Volume I. Louis Berger Group, Inc. Delaware
Department of Transportation Archaeology Series, No. XXX (draft). Dover.
92
Journal of Middle Atlantic Archaeology, Volume 21, 2005
Mueller, Raymond G. and John C. Cavallo
1995
Multiple Working Hypotheses: An Analysis, Pit Houses or Tree Throws. Paper presented at
the 1995 ESAF Conference, Wilmington, Delaware.
Petraglia, Michael, Dennis Knepper, John Rutherford, Philip LaPorta, Kathryn Puseman, Joseph
Schuldenrein, and Noreen Tuross
1998
The Prehistory of Lums Pond: The Formation of An Archaeological Site in Delaware.
Delaware Department of Transportation Archaeology Series No. 155. Prepared for the
Delaware Department of Transportation, Dover, Delaware by Parsons Engineering Science,
Inc., Fairfax, Virginia.
Petraglia, M.D., S.L. Bupp, S.P. Fitzell, and K.W. Cunningham, compilers
2002
Hickory Bluff: Changing Perceptions of Delmarva Archaeology. Delaware Department of
Archaeology Series, No. XXX (draft). Prepared for the Delaware Department of
Transportation, Dover, Delaware by Parsons Engineering Science, Inc., Fairfax, Virginia.
Smith, B. D.
1989
Origins of Agriculture in Eastern North America. Science. 246:1566-1571.
Smith, B. D. (editor)
1992
Rivers of Change: Essays on Early Agriculture in Eastern North America. Smithsonian
Institution Press, London and Washington DC.
Stevens, J Sanderson.
1991
A Story of Plants, Fire and People: The Paleoecology and Subsistence of the Late Archaic
and Early Woodland in Virginia. In Late Archaic and Early Woodland Research in
Virginia: A Synthesis, edited by T.R Reinhart and M.N. Hodges, pp. 185-211. Council of
Virginia Archaeologists, Archaeological Society of Virginia Special Publication No. 23.
Sutton, Ann and Myron Sutton
1985
Eastern Forests. The Audubon Society Nature Guides. Alfred A. Knopf, New York, NY.
Thomas, Ronald A.
1981
Excavations at the Delaware Park Site (7NC-E-41). Delaware Department of
Transportation Archaeological Series No. 11, Dover.